As is well known, fiber optic technology is a rapidly growing field with vastly expanding commercial applicability. As with all technologies, fiber optic technology is faced with certain practical difficulties. In long haul optical transmission systems, optical signal power loss causes unpredictable but significant losses in signal strength. Such losses are caused by a variety of factors, including but not limited to, variations in optical path length, equipment characteristics, environmental conditions, the effects of aging, and so on. In view of these and other factors, it has proven difficult to maintain optical signals at relatively uniform power levels as they pass through the optical transmission system. This is particularly so as the optical signals pass through the optical switches of an optical transmission system.
Another factor which contributes to the development of non-uniform signal power in a group of many optical signal channels is related to the need for continuing signal amplification of the optical signals as they negotiate an optical path through an optical transmission system. Optical amplification is required to avoid expensive electrical signal regeneration over optical paths that can extend over thousands of miles. The chain of amplifiers arranged along an optical path is referred to as a cascade of amplifiers. A problem with such cascades of amplifiers is that optical amplifiers have a strong wavelength dependence on gain. This means that the amplifiers amplify optical signal at some wavelengths more than optical signals at other wavelengths. This and other problems induce non-uniformity in optical signal. More troubling, the effects of amplifier gain non-uniformity increase with each amplifier in the cascade of optical amplifiers along the optical path. This means that as the optical signal passes through each amplifier, the effects of amplifier gain non-uniformity intensify. Therefore, the longer the signal path, the greater the number of amplifiers, and as a result, the greater the degree of gain non-uniformity. In addition, the buildup of optical noise from the amplifier gain peaks can quickly saturate a cascade of amplifiers.
When groups of optical signals (channels) having non-uniform optical power pass through switching nodes a variety of problems can occur. One problem of particular significance is the possibility of misdirecting light from one optical channel onto another optical channel. As result, an input signal (or a portion thereof) from one channel is output into the wrong output fiber. Consequently, a signal intended for one output fiber can be contaminated by signal intended for another output fiber. This phenomenon is referred to herein as “cross-talk”. This problem is magnified to a distressing level in situations where the optical signal power in one or more fibers is significantly greater than the optical power in other fibers. For example, if a first optical beam is 10 dBm more powerful than an adjacent second optical beam, if even a small fraction of the light from the first beam is deflected into the path of the second beam, the signal of the second beam will be corrupted by the cross-talk from the first beam. Moreover, as optical switch size steadily decreases, the margin for error in switching systems also decreases. As a result, in systems with non-uniform optical power levels in the fibers, the likelihood of cross-talk and the resulting problems significantly increases.
One conventional approach for addressing gain non-uniformity problems is through the use of static gain equalization using commercial filters, such as fiber Bragg gratings. In such implementation, the wavelength dependent loss related to the filters corresponds approximately to the wavelength dependent gain from optical amplifiers. However, optical amplifier gain is affected by other factors, such as, input signal level, temperature, and amplifier aging effects. As a result, simple Bragg gratings do not provide a satisfactory solution to gain non-uniformity problems.
Numerous other approaches toward solving signal non-uniformity in optical network applications have been tried. Although some of these approaches work well enough in some situations, each suffers from its own set of limitations. All require the addition of new hardware which introduces new causes for signal loss in the system. Additional new hardware increases system complexity, thereby increasing unreliability. Moreover, these hardware systems all increase cost. Therefore, there are continuing efforts to provide improved methods and apparatus for reducing the effects of non-uniform optical power in optical transmission systems without adding new hardware, without substantially increasing system complexity and unreliability, and without increasing cost. Method and apparatus embodiments constructed in accordance with the principles of the present invention are intended to solve these and other problems.